Population structure of Phytophthora infestans in China – geographic clusters and presence of the EU genotype Blue_13


E-mail: theo.vanderlee@wur.nl


The population structure of Phytophthora infestans in China was studied and three mitochondrial haplotypes (Ia, IIa, IIb) were observed. Genetic analysis with 10 highly informative SSR markers identified 68 different genotypes, including three dominant clonal lineages. In the Chinese P. infestans population, the genotypes were strongly clustered according to their geographic origin. One of dominant clonal lineages was genetically similar to Blue_13, a dominant clonal lineage found in Europe since 2004. This is the first report of Blue_13 outside Europe. Only one mating type (A1) was found in the northern and southeastern provinces, but in southern and northwestern China both mating types were observed. The mating type ratio and SSR allele frequencies indicate that in China the sexual cycle of P. infestans is rare. These results emphasize that the migration of asexual propagules and the generation of subclonal variation are the dominant driving factors of the population structure of P. infestans in China. They may also have implications for the role of monitoring P. infestans populations in potato late blight management strategies in China.


Late blight, caused by the oomycete Phytophthora infestans, is one of the most devastating diseases of potato and tomato. Phytophthora infestans has a heterothallic mating system and the two mating types (A1 and A2) are required for sexual reproduction (Judelson, 1997). The climate in Chinese potato production regions is mostly favourable for P. infestans and potato late blight is currently the most serious threat to potato production in China (Huang et al., 1981). Potatoes in China are grown in four zones based on climate and different agricultural practices. In northern China, potatoes are grown from May to October on large-acreage fields. In middle China they are grown in medium-sized fields and planted twice a year. In western China (Sichuan and Yunnan provinces), they are cultivated on small acreages all year. In southeastern China (Fujian province), potatoes and tomatoes are grown in the field, with potatoes grown mainly in the winter and tomatoes in the spring (Chen & Qu, 2008). Similar to observations in Europe and the USA, potato late blight epidemics in China have recently become more difficult to control. In Europe and the USA this has been attributed to a change in the population structure (Cooke et al., 2009; Hu et al., 2012).

Prior to the 1980s, only the A1 mating type of P. infestans was found outside Mexico. Subsequently, the A2 mating type has been reported from several parts of the world. In China the A2 mating type was first found in northern China in 1996 (Zhang et al., 1996) and later detected in other regions, including Yunnan in southern China close to the Vietnamese border (Zhao & Zhang, 1999). Although both mating types are found in China, to date no evidence of an active sexual cycle has been found based on allele frequencies (Yang et al., 2008).

A new clonal A2 lineage was found in the Netherlands in 2004 (Dr Y. Li, Wageningen University, Netherlands, personal communication) and in the UK in 2005 (Cooke et al., 2009). Isolates with this genotype are highly aggressive, spread rapidly and now dominate P. infestans populations in large parts of Europe, including the UK, Ireland (Cooke et al., 2009) and the Netherlands (Dr Y. Li, personal communication). This clonal lineage, named Blue_13 is considered a new threat to potato production and its spread is intensively monitored in Europe. So far, the Blue_13 lineage has not been reported outside Europe.

The potato agribusiness in China has grown rapidly in recent years (Chen & Qu, 2008). Nationwide transportation of commercial (seed) tubers and potato commodities occurs more frequently and will boom in the coming years with support from the Chinese government. This increases the risks of (i) introducing new pathogen genotypes from abroad, (ii) the spread of more aggressive isolates and (iii) bringing together the two mating types.

Strategies for durable control of potato late blight (PLB) built on integrated pest management (IPM) principles are counteracted by population changes as described above for Europe and the USA. Monitoring of the local P. infestans populations for new virulent genotypes and/or screening for effector variation allows early detection of adaptation within the P. infestans population in a certain region. Monitoring the pathogen population for virulence can enhance the durability of newly introduced host resistance genes. Unfortunately, current knowledge of the Chinese P. infestans population is based on studies that have relied on a few molecular markers and a limited number of P. infestans isolates (Guo et al., 2009, 2010; Yao et al., 2009). The conclusions drawn from these studies are conflicting. Guo et al. (2009) characterized 48 P. infestans isolates, collected in five provinces in northern China during the period 1997–2003. Phytophthora infestans isolates in that study all belonged to the same clonal lineage, as determined by two SSR markers, with some additional subclonal variation, as determined by AFLP markers. However, determination of the virulence spectrum towards Solanum demissum R-genes R1R11 revealed a high diversity within this clonal group (Guo et al., 2009). In contrast, other studies showed more genetic variation among isolates (Yang et al., 2008; Yao et al., 2009). Only Guo et al. (2010) observed migration of P. infestans from neighbouring countries to China. This most recent study analysed 100 isolates collected from 10 provinces in China between 1998 and 2004.

The objective of this study was to assess P. infestans isolates from seven major potato producing regions of China and explore: (i) the current population structure; (ii) the existence of sexual reproduction; (iii) the degree of clonal variation within the population diversity; (iv) migration patterns within the country; and (v) the virulence spectrum of the isolates.

Materials and methods

Sampling of isolates

The isolates were provided by researchers from local universities and research institutes, who visited infected fields reported by farmers and collected infected leaves from fields used in potato evaluation programmes. Sampling intensity varied by year and region, depending on late blight outbreaks. The sampling periods overlapped the local potato season(s) in the different regions, as mentioned in the introduction. In total, 188 isolates were provided originating from seven regions, including 12 provinces, collected between 2004 and 2009 (Table S1; Fig. 1). In addition, a set of 41 isolates from a previous study (Guo et al., 2009) was used for SSR genotyping. The latter set could only be analysed genotypically, as only DNA was available. The seven regions were HLJ (including Heilongjiang and Jilin provinces), IM (Inner Mongolia province), HB (Hebei province), SC (including Sichuan and Guizhou provinces), YN (Yunnan province), FJ (Fujian province) and NW (northwestern China, including Gansu, Shanxi and Ningxia provinces). The number of isolates for each region is shown in Table 1. Each isolate represents one infected field, except for three isolates from a field in Kunming, which were collected from different plants. In addition, it should be noted that Sichuan (in western China) is a mountainous region and although the sample locations on the national map seem nearby, they are often separated by mountain barriers and the travelling distance between sample locations is at least 40 km. Samples from Fujian were collected from potato and tomato. For reference, isolates collected before 2004 from potato in Fujian and Yunnan were included.

Figure 1.

Sampling location map of Phytophthora infestans isolates in China. Dots represent the sampling locations by GPS coordinates; red dots indicate the samples collected for this study between 2004 and 2009; green dots indicate samples from the study of Guo et al. (2009).

Table 1. Summary of Phytophthora infestans isolates by region of collection
RegionaIsolatesMTbHTcMGsd H' e H'/ln(g)e G e G/ge N 1 e E 5 e
  1. a

    HLJ, Heilongjiang and Jilin; IM, Inner Mongolia; HB, Hebei; SC, Sichuan and Guizhou; YN, Yunnan; FJ, Fujian; NW, northwestern China.

  2. b

    Mating type.

  3. c


  4. d

    Multilocus genotypes.

  5. e

    Genotypic diversity analysis H, G, N1: indices of diversity; H '/ln(g), G/g, E5: indices of evenness (Grünwald et al., 2003).

IM52A1IIa 81·430·363·130·064·180·67
HB21A1IIa 41·050·342·380·112·860·74
NW10A1/A2Ia/IIa 51·360·593·120·313·900·73

Isolate purification and long-term storage

Potato leaves with a single lesion were collected from the field and placed in plastic bags or in 9-cm Petri dishes containing 1·5% water agar and incubated until sporulation. Infected leaves were placed underneath tuber slices inside an otherwise empty 9-cm Petri dish. After 5–7 days at 20°C, mycelium emerging from the top of the tuber slice was transferred to pea agar (PA) with ampicillin (Shattock et al., 1990). All isolates in this study were single-zoospore cultures. One isolate per infected leaf was maintained.

For long-term storage the isolates were grown for 1–2 weeks on PA at 20°C. Five agar plugs with fresh mycelium were transferred to a cryotube vial (1·8 mL), after which 1·5 mL 15% sterile dimethylsulphoxide (DMSO) was added. Within an hour, the filled vials were frozen at −80°C for 24 h. After this pre-freezing step, the vials were quickly transferred to a liquid nitrogen storage tank.

Determination of mating types

An agar plug obtained from the edge of an actively growing colony of the test isolate was transferred to one side of a PA-containing Petri dish and a mycelium plug of an A1 tester (isolate VK98014) or an A2 tester (isolate EC3425) was placed on the other side of the Petri dish. Plates were incubated in the dark for 14–21 days at 18°C. After mycelial contact between both colonies was established, the contact zone was monitored for the presence of oospores regularly during 7 consecutive days using a microscope at × 100 magnification. When oospores were found in the Petri dish with the A1 tester isolate the unknown isolate was classified as the A2 mating type and vice versa.

Because of the difficulties in preserving living isolates in hot summers under basic laboratory conditions in the local institutes, only part of the collection survived long enough to perform a mating type test (Table S1).

Preparation of sporangial suspensions

An agar plug obtained from the edge of an active colony growing on PA medium was placed underneath a potato tuber slice inside an otherwise empty 9-cm Petri dish. Inoculated tuber slices were incubated for 6–8 days in a climate chamber (15°C at a light intensity of 12 W m−2 for 16 h day−1). Phytophthora infestans mycelium, growing through the tuber slices, was transferred to a single drop of water, which was placed on the adaxial (upper) side of a potato cv. Bintje leaf, inside a 9-cm Petri dish containing 1·5% water agar (WA). Inoculated leaves were incubated for 6–8 days in a climate chamber using the conditions described above. Following incubation, sporulating potato leaves were rinsed gently in tap water to liberate sporangia. Sporangial suspensions were adjusted to 1–2 × 104 sporangia mL−1 to serve as inoculum for the detached-leaf virulence assays.

Virulence assays

To generate reliable virulence profiles for the isolates, virulence assays were replicated three times. A total of 21 isolates representing all available SSR genotypes and six different geographic origins (Table 1) were selected for virulence assays. This representative selection included isolates from within and among different clonal subgroups in order to study the relationship between genotype and virulence pattern. Virulence phenotypes were determined using a detached-leaf assay employing a combination of Black's and Mastenbroek's differential sets for S. demissum R-genes R1R11 (R1 (Mastenbroek 43154-5), R2 (Mastenbroek 44158-4), R2 (Black 1512c), R3b (Mastenbroek 4642-1), R4 (Mastenbroek 4431-5), R5 (Black 3053-18), R6 (Black XD2-21), R7 (Black 2182ef(7)), R8 (Black 2424a(5)), R9 (Black 2573), R10 (Black 3618ad(1)) and R11 (Black 5008ab(6)) (Black et al., 1953)) and potato cv. Bintje as a susceptible control. Each experiment contained two leaflets per differential × isolate combination per Petri dish and two Petri dishes per replicate experiment. Leaflets were inoculated by spraying them with a sporangial suspension of 2 × 104 sporangia mL−1 of the appropriate P. infestans isolate. Petri dishes containing the inoculated leaflets were incubated at 15°C with a 16-/8-h light/dark regime. Infection severity was assessed per leaflet visually after 1 week of incubation, as described by Flier & Turkensteen (1999).

In planta production of oospores

Ten isolates from different provinces were selected for the in planta crossing test. Sporangial suspensions of an A1 and an A2 mating isolate (1 × 104 sporangia mL−1), produced as described above, were prepared and mixed in equal amounts. This mixture was then cooled for at least 30 min at 4°C to allow liberation of zoospores. Meanwhile, leaflets or leaves of cv. Bintje were placed abaxial (lower) side up in 14-cm Petri dishes containing 1% WA. The mixed sporangial suspension was sprayed onto the potato leaflets using a spraying nozzle (0·5 kg m−2 pressure) until the leaves were completely covered with small droplets. Petri dishes were placed in a plastic tray on wet filter paper and wrapped in a transparent plastic bag to prevent dehydration. The inoculated leaves were incubated at 15°C in the dark for 1 day followed by at least 2 weeks at 15°C and a light intensity of 12 W m−2 for 16 h day−1. To prevent dehydration, the leaves were sprayed with tap water after 1 week. Presence of oospores was assessed visually following incubation using a binocular microscope (Leica) at × 100 magnification.

DNA extraction

Agar plugs of each individual P. infestans isolate were taken from the edge of a 7-day-old actively growing colony on PA and transferred to liquid pea broth. After 3–4 days' incubation at 20°C in the dark, sufficient mycelium was available for DNA extraction. Genomic DNA was isolated from 20 mg lyophilized mycelium using the DNeasy 96 Plant Kit (QIAGEN) following the manufacturer's instructions and eluted in 200 μL ultrapure water. DNA extracts were stored at −20°C until further use.

Haplotype test

Mitochondrial haplotypes were determined using the PCR-RFLP method of Griffith & Shaw (1998). Restriction digestions of the amplified regions P2 (MspI) and P4 (EcoRI) allowed the differentiation of four mitochondrial (mtDNA) haplotypes: Ia, Ib, IIa and IIb.

Microsatellite genotyping

Ten microsatellite markers were used in this study for the genotypic study. Markers used were Pi4B and G11 (Knapova & Gisi, 2002), Pi63 and Pi70 (Lees et al., 2006) and PinfSSR2, 3, 4, 6, 8 and 11 (Li et al., 2010). The forward primers of nine markers and one reverse primer of PinfSSR6 were labelled with VIC, FAM, NED and PET (Applied Biosystems; Table 2). Amplification reactions were as described previously (Li et al., 2010), with some minor modifications. Amplifications were run in a PTC200 thermocycler (MJ Research), with an initial denaturation at 95°C for 15 min, followed by 30 cycles of 95°C for 20 s, 58°C for 90 s and 72°C for 60 s, plus a final extension at 72°C for 20 min. One to two microlitres of the PCR product was added to 1 μL deionized formamide loading buffer and denatured at 92°C for 3 min. The resulting amplification products were sized by capillary electrophoresis on an automated ABI 3730 using the molecular standard GeneScan-500 ROX and scored using GeneMapper v. 3.7 software (Applied Biosystems). To facilitate scoring and the setup of an international database, a total of 11 reference isolates were used, including eight reference isolates for SSR analysis (T30-4, 80029, 88133, VK1.4, 90128, IPO-0, IPO428-2, VK98014; Li et al., 2010) from Europe and three Blue_13 genotype variants (NL05246, NL05238 and NL05147). NL05238 has the typical genotypic profile of isolates that belong to the Blue_13 clonal lineage analysed with the same SSR markers (provided by Dr D. Cooke, The James Hutton Institute, UK).

Table 2. Profile of SSR markers
IDSize rangeLabelPrimer sequence (5′ to 3′)No. allelesPICa H E b H O c P-value (HWE)d
  1. a

    Polymorphism information content.

  2. b

    Expected heterozygosity.

  3. c

    Observed heterozygosity.

  4. d

    Significance of deviation from Hardy–Weinberg equilibrium.













SSR 2165–180PET



SSR 3254–274NED



SSR 4280–3056-FAM



SSR 6230–260VIC



SSR 8256–2746-FAM



SSR 11325–360NED




Data analysis

Based on the SSR data, pairwise comparisons of genetic distances were calculated using Nei's formula (Nei, 1972). A genetic distance matrix was established and subsequently used to construct a dendrogram based on the neighbour-joining (NJ) clustering procedure. The SSR data were scored by a binary representation of the presence (1) or absence (0) of alleles and analysed using the phylogenetic software package treecon for Windows v. 1.3b (Van de Peer & De Wachter, 1994). Bootstrapping was performed by using 1000 bootstraps. The phylogenetic tree was generated using genotypes of P. infestans (Table S1) and three Dutch isolates.

To estimate diversity, polymorphism information content (PIC), number of alleles per locus (Na), observed heterozygosity (HO), expected heterozygosity (HE) and deviations from the Hardy–Weinberg equilibrium (HWE) were determined. The significance of the deviations from the HWE were calculated with exact P-values estimated using the Markov chain algorithm with 10 000 dememorization steps, 100 batches and 1000 iterations using genepop v. 4.0 (Raymond & Rousset, 1995). To examine the distribution of genetic variation, analysis of molecular variance (amova; Excoffier et al., 1992) was performed using winamova (Excoffier et al., 1992). Raw data were preprocessed by amova-prep (Miller, 1998) to generate the amova input files. amova analyses were then performed with significance tests for 1000 permutations to determine how the genetic diversity was partitioned within and between populations. Multilocus genotypic diversity was estimated using Stoddart & Taylor's index G (Stoddart & Taylor, 1988) and Shannon & Wiener's H′ (Shannon, 1948).

To examine the genetic structure of Chinese P. infestans isolates, the clustering program structure v. 2.2 (Falush et al., 2003) was run after clone correction (one representative isolate for each genotype). The number of clusters (K) was specified from 1 to 15, five independent runs were conducted to assess the consistency of the results across runs, and all runs were based on 100 000 iterations after a burn-in period of 100 000 iterations. The statistic ΔK, which indicates the highest level hierarchical structure in the population (Evanno et al., 2005), was calculated. To perform ΔK calculations, the likelihood from each of five structure runs from each K was randomly assigned into one of five groups, each containing a single likelihood from each K. To validate the genetic structure, principal component analysis (PCA) using NTSYS-pc v. 2.0 software (Rohlf, 1987) was conducted to construct plots of the most significant axes for grouping using the default settings.

Identification of subclones and naming system

The set of SSR markers used was highly informative, allowing the identification of subclonal lineages within a population. Subclones were defined as nearly identical genotypes only separated from the main lineage by the absence of specific alleles in some SSR loci, which could be the result of asexual mitotic recombination (Chamnanpunt et al., 2001) or chromosomal deletions (van der Lee et al., 2001). In addition, SSR loci G11 and PinfSSR4 are extremely variable, resulting in subclonal variation showing new alleles within an otherwise identical genetic background (Cooke et al., 2009; Li et al., 2010). The naming of the clonal and subclonal genotypes started with the two-letter country abbreviation (CN), followed by the ranking number of the clonal lineage in the study. For example, one main clonal lineage found in this study was CN01. The subclonal genotypes within this clonal were named CN01_01, CN01_02, … and so on.


Mitochondrial haplotype analysis

Three mtDNA haplotypes (Ia, IIa, IIb; Fig. 1 and Table 1) were found among the Chinese isolates in this study, and the occurrence of each was strongly associated with the sampling origin of the isolates (P-value of χ2 test < 0·0001; Figs 1 and 2). Nearly all the isolates (59 of 60) derived from the northern part of China, i.e. Heilongjiang, Inner Mongolia and Hebei, had the IIa haplotype; only one isolate (HLJ05-NL1 from Heilongjiang) had the mtDNA haplotype Ia. All isolates derived from Sichuan and Yunnan had the Ia haplotype. All isolates collected in Fujian province had the IIb haplotype. mtDNA haplotype Ib (US-1 clonal lineage) was not found in this study.

Figure 2.

Genotype frequency of Chinese Phytophthora infestans isolates. CN01–CN03 are main clonal lineages; the suffixes _01, _02, etc. indicate subclonal genotypes. Sampling regions: NW, northwestern China; FJ, Fujian; YN, Yunnan; SC, Sichuan and Guizhou; HB, Hebei; IM, Inner Mongolia; HLJ, Heilongjiang and Jilin.

Profile of SSR markers

The 10 SSR markers yielded 44 alleles from the 229 isolates, with an average of 4·4 alleles per locus. For Pi70, no polymorphism was found among Chinese isolates. For the polymorphic markers, an average of 4·9 alleles was detected for each locus. The mean PIC value was 0·58. G11 and PinfSSR4 were the most informative SSRs, with the highest PIC value (0·76). The mean expected heterozygosity and observed heterozygosity were 0·43 and 0·65, respectively. A complete list of loci and their variability is shown in Table 2. Except for G11, all loci deviated significantly from the HWE expectations (< 0·05), as observed with a nonrandom mating organism. Linkage disequilibrium between pairs of loci was detected, but was not consistent across geographic populations. Tri-allelic isolates were observed for six markers (G11, Pi4B, PinfSSR3, PinfSSR4, PinfSSR8 and PinfSSR11). The highest frequency of tri-allelic isolates was 38% for G11; three markers (Pi4B, PinfSSR8 and PinfSSR11) showed only a low tri-allelic frequency (<6%).

Genetic variation by region

SSR genotyping of the 229 P. infestans isolates resulted in 68 distinct genotypes (Fig. 2), 48 of which represented subclonal variations within the three dominant clonal lineages. Careful inspection of the genotypes revealed that these subclonal variations comprised losses of alleles for a particular locus or the presence of rare alleles in the hypervariable loci PinfSSR4 and G11. The dominant genotype, CN01_01 (Fig. 2), was found in four out of seven regions (HLJ, IM, HB and NW) located in the north and northwest of China, (Fig. 3). Its subclones CN01_07 and CN01_08 were also identified at multiple locations in northern China. A second dominant genotype, CN02_01, was found in three regions (NW, YN and SC). The other 65 genotypes, including 25 genotypes belonging to clonal lineage CN03, were unique to the province from which they were collected. Genotypic variation within the seven regions varied from 5 (NW) to 24 (FJ) genotypes (Fig. 2; Table 1). Among all seven regions YN and FJ showed the most variable genotypic diversity indices and evenness (Table 1).

Figure 3.

Clustering tree of Chinese Phytophthora infestans isolates based on SSR data. Genetic distance estimation was performed according to Nei & Li (1979) and clustering was performed using neighbour-joining. Bootstrap values in > 50% from 1000 replicate trees are shown at the nodes. The dark lines on the left side indicate the main clonal lineages. Mating types and haplotypes are shown on the right. The tree was rooted with VK1.4, a US-1 genotype originating from the Netherlands.

To examine a potential substructure within the Chinese population, the isolates were divided into seven subpopulations based on sampling regions. In the amova analysis, 26·1% of the variation was found within subpopulations, while the variation between subpopulations was significantly (< 0·001) higher at 73·9%. To visualize the genotypic relations, a similarity tree was constructed using only the SSR data by neighbour-joining analysis based on Nei's genetic distances. The phylogram (Fig. 3) was rooted with the US-1 genotype from the Netherlands (VK1.4, Ib). The genotypes grouped into four distinct clusters. A clear pattern emerged when the phylogeny tree was linked with the geographic regions. Isolates of cluster I had the A1 mating type and the mtDNA haplotype IIa with the exception of isolate HLJ05-NL1, which had the Ia haplotype. Cluster I contained all 100 isolates derived from northern China with the addition of seven isolates from NW and SC. The clonal lineage CN01 was identified in this cluster. Cluster II consisted of isolates from Yunnan and Sichuan that had the A1 mating type and the Ia haplotype. Two isolates from Yunnan (YN07-501805 and YN07-401204) had a distinct genotype, with different alleles at five loci compared with other Yunnan isolates. One genotype in cluster II, representing four Sichuan isolates, was also clearly distinct. The two isolates from Sichuan in this cluster (SC07-M0102 and SC07-M0103) were sampled in the same county.

Cluster III consisted of isolates from Sichuan and a few isolates from NW with the mating type A2 and the haplotype Ia. Nearly all isolates (68 of the 70) belonged to the same clonal lineage (CN02). The isolates from Sichuan province were collected from different years and fields in the mountain-rich areas. Three Dutch isolates representing subclones of the Blue_13 clonal lineage (NL05246, NL05238 and NL05147) were included in the analysis and dispersed among the isolates of the CN02 clonal lineage, demonstrating that the CN02 clonal lineage is part of the Blue_13 clonal lineage. Cluster IV groups all isolates from Fujian derived from both potato and tomato. All isolates from cluster IV had the IIb haplotype and the A1 mating type. The isolates collected in Fujian showed a high level of subclonal variation. Only five SSRs (Pi4B, Pi63, Pi70, PinfSSR2 and PinfSSR6) showed the same profile in all isolates, the other five polymorphic SSRs showed 15 different alleles, 11 of which had less than 11% variation. Eight isolates were obtained from potato plants, whereas the rest of the Fujian isolates originated from tomato. In Fujian, the P. infestans isolates collected from potato and tomato were genetically similar. The isolates collected from potato were not clustered in a particular clade, but were found to be dispersed among the majority of isolates collected from tomato.

Population structure

Genetic structure was analysed with the model-based clustering algorithm implemented in the structure v. 2.2 software. To avoid redundancy in the collection, the data were clone-corrected, keeping only one isolate when more than one isolate was identified with the same SSR genotype at all loci. In total, 68 genotypes were included in the analysis. The Evanno et al. (2005) correction of the structure output was used (Fig. S1). The first peak of ΔK, for = 3, corresponded to the presence of three main groups. One structure group included Fujian isolates (shown in blue in Fig. S1), while other isolates were divided into two groups coloured in red (isolates in clusters I and II) and green (isolates in cluster III). The results indicated that almost all isolates (>90%) had a high membership of their own group (>90%), indicating that there had been little or no gene flow between these three groups. To validate the genetic structure identified by phylogeny and structure, PCA was conducted. Plotted on the first two dimensions of PCA, the four clusters previously detected by the phylogeny analysis (Fig. 3) were adequately separated from each other (Fig. 4).

Figure 4.

Two-dimensional plot (= PC1, = PC2) of a principal components analysis of Chinese Phytophthora infestans isolates. Sampling regions: HLJ, Heilongjiang and Jilin; IM, Inner Mongolia; HB, Hebei; SC, Sichuan and Guizhou; YN, Yunnan; FJ, Fujian; NW, northwestern China.

Mating analysis

Mating type was determined for 114 viable isolates. All isolates produced oospores either with the A1 or A2 tester and no selfing isolates were found. Both mating types were found among the sampled isolates. The A1 mating type predominated and was found in all sampled provinces. All isolates from Heilongjiang, Inner Mongolia, Hebei and Fujian had the A1 mating type. In Yunnan, Sichuan, Ningxia and Shanxi, two mating types were found. However, the frequency of the A2 mating type (10%, one of 10) in Yunnan strongly contrasted with the frequency in Sichuan (91%, 21 of 23). To confirm the potential compatibility of the two mating types in China, nine crossing combinations were performed. These crosses included both isolates originating from the same province as well as isolates from the north and south of China (Table 3). High densities of oospores were observed in all mating combinations tested (Table 3).

Table 3. In planta crossing experiments between Chinese Phytophthora infestans isolates
A1 isolateYN07-501805YN07-500504HLJ06-HH13IM07-6·1YN05-LSX18SC07-M0102SC07-M0102SC07-M0102SC07-M0102
A2 isolateYN05-d5-2-2YN05-d5-2-2YN05-d5-2-2YN05-d5-2-2YN05-d5-2-2YN05-d5-2-2SC07-S0902SC07-X0103SC07-Pt0109

Virulence analysis

The virulence spectrum was determined for a selection of 21 isolates representing different SSR genotypes (Table 4). With the exception of the R2 differential, which showed high variation among duplicated virulence assays, all virulence results were consistent. Because of variable results of the R2 differential in duplicate tests, these results were excluded. Isolates selected from cluster I and cluster II, including representatives from Inner Mongolia, Heilongjiang, Hebei, Yunnan and one isolate from Sichuan (SC07-M0102), had the same virulence spectrum (1, 3b, 4, 7, 10, 11) with no virulence to R5, R6, R8 and R9. Isolates from cluster III, comprising CN02 subclones, showed complex virulence patterns (1, 3b, 4, 6, 7, 10, 11) with occasional virulence on potato lines carrying R5, R8 and R9. Isolate SC07-S0902 (CN02_01) had the most complex virulence pattern (1, 3b, 4, 5, 6, 7, 8, 9, 10, 11). Isolate YN-d5-2-2 from Yunnan, but grouped in cluster III (CN02_08), also had a complex virulence pattern (1, 3b, 4, 6, 7, 10, 11). Fujian isolates (cluster IV) showed a less complex virulence pattern (1, 3b, 4, 7). The virulence spectrum differed strongly for each region and the virulence pattern correlated strongly with the genotyping background.

Table 4. Virulence assay on a selection of 21 Chinese and three reference Phytophthora infestans isolates using a differential set of 11 potato clones and the susceptible potato cv. Bintje. Clusters I, II, III and IV are based on the phylogenetic tree of Chinese isolates shown in Figure 4
Cluster IpotatoIMIM 07-6.1+++++++++
potatoIMIM 07-3.4+++++++++
potatoIMIM 07-6.5+++++++++
potatoIMIM 07-7.3+++++++++
potatoHLJHLJ 06-HH13+++++++++
potatoHLJHLJ 06-KS622+++++++++
Cluster IIpotatoYNYN07-500504+++++++++
Cluster IIIpotatoYNYN05-d5-2-2++++++++++
Cluster IVtomatoFJFJ06-t77+++++++

Reference Chinese isolates for three main clonal lineages

To name the clonal lineages consistently following a standard nomenclature for future studies, a reference isolate was specified for each main clonal lineage (Table S1).


In this study, the Chinese P. infestans population consisted of multiple genotypes and a clear genetic substructure associated with the origin of the isolates was identified. This study differentiated more subclones in northern China than a previous study (Guo et al., 2009), which could be attributed to the high resolution of the SSR markers and the extended number of isolates tested. The analysis represents the first population study for P. infestans in China, using a high-resolution set of 10 SSR markers and a wide sampling strategy, including isolates from north, south, east and west China.

One surprising finding in this study was that the A2 clonal lineage Blue_13 is present in China. This clonal lineage dominates the P. infestans population in the UK and other European countries and is highly aggressive and virulent (Cooke et al., 2009). Furthermore, this study showed that the Chinese Blue_13 clonal lineage (CN02) contains several novel genetic variants. CN02 has firmly established itself in Sichuan during recent years. In Sichuan province it was found from 2007 and this study also identified one isolate of this lineage from Yunnan that was collected in 2005. With the current data it was not possible to trace back when CN02 arrived in China and where it came from. Sichuan and Yunnan are mountainous areas with favourable weather conditions for year-round potato planting. Potato is grown on small acreages using farmer-saved seed tubers, which may provide a special niche for P. infestans. Official reports do not document any import of seed potatoes from abroad since 2000. This then raises a question about the origin and spread of CN02 in China. Was it present earlier in Sichuan or other Chinese regions, or did it perhaps migrate to China? Fortunately, adequate data sets allowing these comparisons could be available in the near future via an internationally shared database using the same set of SSR markers. A European-scale database developed by the European Concerted Action on Blight with 37 European partners (http://www.euroblight.net) is expanding to a broader international scale.

Isolates from Fujian formed a clearly distinct group in the PCA with a wide range of subclonal variation, indicating that P. infestans in Fujian province has been isolated from other Chinese regions for a long time. The Fujian P. infestans population did not have different subpopulations associated with potato and tomato. Potatoes transported from other Chinese provinces to Fujian could result in the migration of more aggressive potato isolates, which could then aggravate late blight problems in Fujian. It is hypothesized that the Fujian population is an older, less aggressive clonal lineage, which has survived as a result of regional isolation and which may be primarily adapted to tomato.

In a previous study (Zhang et al., 2001), the frequency of the A2 mating type was much higher in northern China (Inner Mongolia and Hebei) than in southern China (Sichuan and Yunnan). However, in this study, no A2 isolates were found in northern China, corroborating the reported decline of the A2 mating type in China (Li et al., 2009) and the absence of the A2 mating type in northern China (Guo et al., 2009). Although the presence of both mating types was observed by Zhang et al. (1996), only Zhao et al. (2001) reported the presence of oospores under field conditions in a single sample in 2000. Crossings in planta in the present study yielded high densities of oospores, indicating that Chinese isolates could initiate sexual reproduction, but further progeny germination and fitness tests are required to check whether any postzygotic mating barriers are present or not. Even if these barriers do not exist, the chance of physical contact between the two mating types is low as both mating types were only found together in some parts of China, and always with a big disparity in the mating type ratio. In addition, the high level of linkage disequilibrium indicates a strong bias towards clonal spread in China.

The virulence pattern of 21 representative P. infestans isolates on the R1R11 differential set of potato lines showed a high level of complexity with regional differences. The virulence pattern fitted well with the SSR genotyping, in contrast to previous studies (Knapova et al., 2002; Guo et al., 2009). The resolution of the SSR markers in the present study was much higher than in these previous studies, and the markers allowed the clear distinction of subgroups. Because virulence is genetically inherited (van der Lee et al., 2001) one would expect that with enough isolates and sufficient resolution in genotyping, the clonal isolates would show identical or at least highly similar virulence patterns. Although the correlation between a specific genotype and its virulence profile cannot be expected to be absolute, the relative stability of virulence of a clonal lineage could help to predict the virulence of new isolates that belong to previously identified clonal lineages.

To conclude, the 10 SSR markers used in this study have uncovered a spatial structure consistent with the model of a metapopulation for the Chinese P. infestans population. The clonal lineage CN01, identified by Guo et al. (2009), is still dominant in China. Using the SSRs, it is now possible to combine and compare genotyping data from different laboratories. The results show that such a global genotyping approach is essential to study the pandemic spread of the Blue_13 clonal lineage. This is apparently the first report of Blue_13 outside Europe. SSR markers provide new possibilities for tracking and tracing the worldwide migration pattern of this aggressive strain. The survey in this study provides a good basis for future studies to understand the dynamics of the national population structure in China. A nationwide survey with deeper sampling could explore the spatial structure and gene flow of P. infestans in China in more detail. Improved throughput for genotyping large numbers of lesions may be achieved by the storage of DNA from foliar lesions on FTA cards (Whatman). In addition, the results indicate that the transportation of potato and tomato commodities, in particular the transportation of seed potatoes from Sichuan to northern or western parts of China and vice versa, may increase the risk of bringing together the two mating types, and hence the possibility of sexual reproduction. This potential sexual reproduction will boost genetic variation and result in the formation of persistent oospores, which could have a dramatic impact on the epidemiology of late blight in China. Monitoring of P. infestans will be necessary to trace the movement and diversification of the main clonal lineages and to identify new dominant genotypes, newly introduced lineages, or recombinant genotypes.


The authors want to thank the Institute of Vegetables and Flowers (Beijing, China) and the National Natural Science Foundation of China (31000738) for financial support, members of the Chinese Late Blight Initiatives for supplying late blight samples, Trudy van den Bosch, Petra van Bekkum and Tineke Veenendaal for excellent technical support and Dr David Cooke (The James Hutton Institute, UK) for providing the genotypic profile of the Blue_13 clonal lineage.